- Email: [email protected]
Marked 25-hydroxyvitamin D deficiency is associated with poor prognosis in patients with alcoholic liver disease
Research Article
Marked 25-hydroxyvitamin D deficiency is associated with poor prognosis in patients with alcoholic liver disease Eric Trépo1,2, , Romy Ouziel1,2, , Pierre Pradat3,4,5, Yukihide Momozawa6, Eric Quertinmont2, Christine Gervy7, Thierry Gustot1,2, Delphine Degré1,2, Vincent Vercruysse2, Pierre Deltenre8, Laurine Verset9, Beatrice Gulbis7, Denis Franchimont1,2, Jacques Devière1,2, Arnaud Lemmers1,2, Christophe Moreno1,2,⇑ 1
Department of Gastroenterology, Hepatopancreatology and Digestive Oncology, Erasme Hospital, Université Libre de Bruxelles, Brussels, Belgium; 2Laboratory of Experimental Gastroenterology, Université Libre de Bruxelles, Brussels, Belgium; 3Hospices Civils de Lyon, Hôpital de la Croix-Rousse, Lyon, France; 4INSERM, U1052, Lyon, France; 5Université Lyon 1, Lyon, France; 6Unit of Animal Genomics, Groupe Interdisciplinaire de Génoprotéomique Appliquée (GIGA-R) and Faculty of Veterinary Medicine, University of Liège (B34), Liège, Belgium; 7 Department of Clinical Chemistry, Erasme Hospital, Université Libre de Bruxelles, Brussels, Belgium; 8Service d’Hépato-Gastroentérologie, Hôpital de Jolimont, Haine-Saint-Paul, Belgium; 9Department of Pathology, Erasme Hospital, Université Libre de Bruxelles, Brussels, Belgium
Background & Aims: Vitamin D deficiency has been frequently reported in advanced liver disease. However, its influence on alcoholic liver disease (ALD) has been poorly elucidated. We investigated the association of vitamin D with clinical, biological, and histological parameters and survival in ALD patients. Furthermore, we explored the effect of vitamin D treatment on ALD patient peripheral blood mononuclear cells (PBMCs), and in a murine experimental model of ALD. Methods: Serum levels of 25-hydroxyvitamin D [25(OH)D] were determined in 324 Caucasian ALD patients and 201 healthy controls. In vitro experiments on vitamin D pre-treated PBMCs evaluated TNFa production by ELISA in culture supernatants. Mice were submitted to an ethanol-fed diet and some of them were orally supplemented three times per week with 1,25(OH)2D.
Keywords: Vitamin D; Genetic factors; Alcoholic liver disease; Inflammation, TNFa.
Results: Severe deficiency in 25(OH)D (<10 ng/ml) was significantly associated with higher aspartate aminotransferase levels (p = 1.00 103), increased hepatic venous pressure gradient (p = 5.80 106), MELD (p = 2.50 104), and Child-Pugh scores (p = 8.50 107). Furthermore, in multivariable analysis, a low 25(OH)D concentration was associated with cirrhosis (OR = 2.13, 95% CI = 1.18–3.84, p = 0.013) and mortality (HR = 4.33, 95% CI = 1.47–12.78, p = 7.94 103) at one year. In addition, in vitro, 1,25(OH)2D pretreatment decreased TNFa production by stimulated PBMCs of ALD patients (p = 3.00 103), while in vivo, it decreased hepatic TNFa expression in ethanolfed mice (p = 0.04). Conclusions: Low 25(OH)D levels are associated with increased liver damage and mortality in ALD. Our results suggest that vitamin D might be both a biomarker of severity and a potential therapeutic target in ALD. Ó 2013 European Association for the Study of the Liver. Published by Elsevier B.V. All rights reserved.
Received 6 July 2012; received in revised form 14 March 2013; accepted 15 March 2013; available online 1 April 2013 ⇑ Corresponding author. Address: Erasme Hospital, Department of Gastroenterology, Hepatopancreatology and Digestive Oncology. Université Libre de Bruxelles. Route de Lennik, 808 1070 Brussels, Belgium. Tel.: +32 2 555 37 12; fax: +32 2 555 46 97. E-mail address: [email protected] (C. Moreno). These authors contributed equally to this work. Abbreviations: HCC, hepatocellular carcinoma; ALD, alcoholic liver disease; BMI, body mass index; AC, alcoholic cirrhosis; NAFLD, non-alcoholic fatty liver disease; CHC, chronic hepatitis C; 25(OH)D, 25-hydroxyvitamin D; 1,25(OH)2D, 1,25-dihydroxyvitamin D; GWAS, genome-wide association studies; PBMCs, peripheral blood mononuclear cells; ALT, alanine aminotransferase; AST, aspartate aminotransferase; AH, alcoholic hepatitis; GGT, gamma-glutamyl transferase; SBP, spontaneous bacterial peritonitis; HRS, hepatorenal syndrome; HVPG, hepatic venous pressure gradient; SNP, single nucleotide polymorphism; ConA, concanavalin A; ELISA, enzyme-linked immunosorbent assay; TNFa, tumor necrosis factor-a; RT-PCR, reverse transcriptase polymerase chain reaction; HPRT, hypoxanthine-guanine phosphoribosyltransferase; EtOH, ethanol; Ctrl, control; MELD, model for end-stage liver disease; OR, odds ratio; CI, confidence interval; HR, hazards ratio; HS, healthy subjects.
Introduction Nearly 4% of worldwide mortality is related to excessive alcohol consumption [1]. More specifically, it is a leading cause of cirrhosis and hepatocellular carcinoma (HCC), which may ultimately requires liver transplantation [2]. Alcoholic liver disease (ALD) and its complications increase in proportion to daily ethanol consumption [3]. Moreover, clinical factors (older age, female sex, body mass index [BMI]) have been associated with a higher prevalence of alcoholic cirrhosis (AC) [4]. In addition, a genetic variant in the PNPLA3 (rs738409 C>G) gene was recently linked to AC in Caucasians [5,6]. Nonetheless, new diagnostic biomarkers and therapeutic targets are urgently needed when confronting alcohol-related liver injury [7].
Journal of Hepatology 2013 vol. 59 j 344–350
JOURNAL OF HEPATOLOGY Vitamin D deficiency has been associated with various skeletal and non-skeletal chronic illnesses [8]. In particular, it has also long been reported in various chronic liver diseases, especially at the cirrhosis stage [9]. More recent studies observed that low vitamin D serum levels were associated with increased histological liver damage in non-alcoholic fatty liver disease (NAFLD) [10] and in chronic hepatitis C (CHC) [11–13]. Whatever its origin (via skin production, diet or dietary supplements), vitamin D must undergo two hydroxylations in order to become biologically active: first in the liver, generating 25-hydroxyvitamin D [25(OH)D], the most relevant indicator of a patient’s vitamin D status; and then in the kidney, as 1,25-dihydroxyvitamin D [1,25(OH)2D]. In addition to its role in bone metabolism and calcium homeostasis, 1,25(OH)2D, the active form of vitamin D, has antitumoral activity, antifibrogenic effects, and anti-inflammatory properties [9]. The aim of the present study was to investigate the association of 25(OH)D deficiency with histological damage, portal hypertension and its related complications, liver function, and mortality in ALD. In addition, since the immunomodulating role of vitamin D has been established, we studied the influence of 1,25(OH)2D treatment upon inflammation in ALD patient peripheral blood mononuclear cells (PBMCs) in vitro, and we evaluated the effects of 1,25(OH)2D oral supplementation in a murine experimental model of ALD in vivo.
For each patient, the date of inclusion was the date of the first liver biopsy or, in its absence, the first blood sampling at the time of admission. Gender, age, BMI, past history of diabetes mellitus, daily alcohol intake, presence of complications of portal hypertension (ascites, hepatic encephalopathy HRS, SBP or bleeding), serum 25(OH)D bilirubin, albumin, prothrombin, gammaglutamyl transferase, ALT, and AST levels were recorded at inclusion. All patients were followed up after discharge from the liver unit. Thus, they were scheduled for new evaluations at a minimum of 6-month intervals, or sooner in case of more severe disease. In addition to conventional monitoring, HCC screening included liver ultrasonography and serum AFP level assessment. Follow-up ended at the date of death or liver transplantation (the patient was considered alive), or at the last recorded visit or information gathering. This date was set as the final time limit for upgrading the patient file using our hospital computerized database. Patients transplanted or lost to followup were included in the analysis and censored at the date of transplantation or last recorded information. In addition, 201 healthy Caucasian social workers from the same geographic area (Brussels area) as ALD patients were included as controls. They had neither clinical nor biochemical evidence (routine blood screening) of liver disease, nor evidence of other major pathological conditions. This group served as a reference for vitamin D levels. None of the patients or healthy controls included in our study used vitamin D supplements or medication known to affect the serum vitamin D concentration. Written informed consent was obtained from all patients and this study was approved by the local ethics committee. Patient and healthy control characteristics are summarized in Table 1.
Biological variables assessment Biological parameters were measured concomitantly with the liver biopsy procedure or at the time of admission. Quantitative assessment of serum total 25(OH)D levels (25-hydroxyvitamin D2 and D3) was performed by direct competitive chemiluminescence immunoassay using commercially available kits from Diasorin on a LiaisonÒ (Diasorin, Stillwater, MN, USA).
Patients and methods Patients A total of 324 ALD patients were recruited prospectively between January 1, 2002 and May 31, 2011 at Erasme Hospital, Brussels, Belgium. According to the ALD definition, these unrelated Caucasian patients had a history of excessive alcohol intake of >30 g/day (76% drank more than 80 g/day), together with increased alanine aminotransferase (ALT) and aspartate aminotransferase (AST) or suspected AC [2]. The diagnosis of cirrhosis was based on liver biopsy and/ or an unequivocal clinico-biochemical profile, a highly suggestive endoscopic exam (e.g., esophageal varices with or without a previous episode of bleeding) or typical imaging findings in agreement with current guidelines [14]. Seventyeight percent of these ALD patients had cirrhosis. The diagnosis of alcoholic hepatitis (AH) was based on both compatible histological signs and a history of recent jaundice onset, with serum bilirubin >5 mg/dl [15]. The presence of HCC was based on computed tomodensitometry and/or magnetic-resonance imaging and/or guided liver biopsy [16]. The diagnosis of spontaneous bacterial peritonitis (SBP) was based on an ascitic fluid neutrophil count P250 cells/ mm3, and that of hepatorenal syndrome (HRS) was in agreement with current guidelines [17]. Patients with other causes of liver disease, including viral or autoimmune hepatitis, hemochromatosis or co-infection with human immunodeficiency virus, were excluded.
Histological assessment Ninety-seven percent of all ALD patients underwent liver biopsy. The percentage of steatosis was determined among the total number of hepatocytes, and fibrosis was assessed according to the Brunt score [18], using Sirius Red staining. Diagnosis of AH was established according to current guidelines [15].
Hepatic venous pressure gradient assessment (HVPG) Using a straight catheter, measurement of HVPG was performed concomitantly with transjugular biopsy in 91% of ALD patients for whom cirrhosis was suspected. This indication was based on the excellent correlation between HVPG and portal hypertension, as was the high predictive value of cirrhosis decompensation [19]. HVPG was calculated as the difference between wedge hepatic venous pressure and free hepatic venous pressure [19]. Each measurement was repeated three times.
Table 1. Characteristics of patients (n = 324) and healthy controls (n = 201).
Female sex, % Age, yr BMI, kg/m2 BMI >25 kg/m2, % 25(OH)D, ng/ml ALT, IU/L AST, IU/L GGT, IU/L Cirrhosis, % AH, %
ALD patients 35.2 54.3 ± 9.4 26.0 ± 5.3 54.3 9.6 (5.0-16.2) 34 (24-49) 55 (38-92) 143.5 (65.3-308) 78.4 20.8
Controls 50.2 43.8 ± 9.7 25.2 ± 5.4 45.5 19.0 (13.5-25.0) 15 (11-21) 24 (19-29) 21.3 (14.4-32.4) -
p value 1.00 x 10-3 1.13 x 10-30 0.084 0.057 4.75 x 10-24 8.51 x 10-45 1.34 x 10-53 1.61 x 10-54 -
BMI, body mass index; 25(OH)D, 25 hydroxyvitamin D; ALT, alanine aminotransferase; AST, aspartate aminotransferase; GGT, gamma-glutamyl transferase; AH, alcoholic hepatitis. Variables with normal distribution are presented as mean ± standard deviation. Skewed variables are presented as median (interquartile range).
Journal of Hepatology 2013 vol. 59 j 344–350
345
Research Article Genotyping DNA was isolated using the phenol chloroform method from a whole blood sample collected concomitantly with liver biopsy. Twenty nanograms of DNA, TaqMan assays (Applied Biosystems, Foster City, CA, USA) and the LightCyclerÒ 480 system (Roche Diagnostics, Mannheim, Germany) were used to genotype PNPLA3 (rs738409, assay ID C_7241_10). Peripheral blood mononuclear cells (PBMCs) PBMCs from 8 healthy volunteers and 8 consecutive AC patients admitted to our institution were freshly isolated using a Ficoll gradient. Then, cells were seeded on 96-well culture plates for 1 h treatment with 107 M 1,25(OH)2D (Sigma Aldrich) or ethanol 0.1% as vehicle control. Thereafter, pre-treated PBMCs were stimulated with concanavalin A (ConA; 2.5 lg/ml) (Sigma Aldrich). Cell-free supernatants were collected 24 h later and stored at 20 °C until the enzymelinked immunosorbent assay (ELISA). Immunoassays An ELISA commercially available kit (Duoset, R&D systems) was used to quantify human tumor necrosis factor a (TNFa). In vivo experiment Eight-week old female C57BL/6 mice (Charles River, Brussels, Belgium) were submitted to the Lieber-DeCarli diet for 6 weeks following a previously published protocol [20]. Some of the mice from the ethanol-fed group (EtOH) were orally supplemented three times per week with 1,25(OH)2D (50 ng; Sigma Aldrich) (EtOH + 1,25(OH)2D; n = 18). Pair-fed (Ctrl; n = 7) and EtOH (n = 22) groups received PBS as vehicle control (protocol detailed in the supplementary section). At the end of the experiment, mice were killed and livers were excised, weighed and frozen at 80 °C, and serum levels of ALT were measured with a commercially available kit (Roche/Hitachi). To explore the potential impact on hepatomegaly (a marker of steatosis) in mice, we calculated a liver/mouse weight ratio as follows: liver/mouse weight ratio = (liver weight/mouse weight) 100. Quantitative reverse transcriptase polymerase chain reaction (RT-PCR) Frozen murine liver samples were first homogenized in lysis solution with a MagNalyser (Roche Diagnostics). Then, total RNA was extracted with the High Pure RNA tissue kit (Roche Diagnostics) according to the manufacturer’s protocol. Reverse transcription and quantitative PCR carried out on murine livers for TNFa and hypoxanthine-guanine phosphoribosyltransferase (HPRT) are detailed in the Supplementary data. The sequences of PCR probes and primers are available in Supplementary Table 1. Statistical analysis 25(OH)D levels are categorized throughout the manuscript using a cut-off of 10 ng/ml for ALD patients and 20 ng/ml for healthy controls. These thresholds were chosen because they were highly similar to the median in both populations (Table 1) and also reflected admitted clinical cut-offs for vitamin D deficiency (<20 ng/ml) and severe deficiency (<10 ng/ml) [21]. We tested the association between 25(OH)D concentrations and seasonal timing of sample collection. Therefore, we compared patients who underwent blood sampling during summer and autumn to those recruited during months with lower sun exposure (winter-spring). Categorical variables were studied using the two-sided Chi-square test, the Chi-square test for trend or the twosided Fisher’s exact test, whereas quantitative variables were assessed using analysis of variance or non-parametric Kruskall-Wallis or Mann-Whitney tests when appropriate. We conducted univariate and multivariable logistic regression to investigate whether 25(OH)D low levels were associated with cirrhosis after adjustment for other clinical (age, sex, BMI, heavy alcohol consumption, and seasonal timing of sample collection) [3,4] and genetic (rs738409 variant in PNPLA3 gene) [5,6] risk factors for AC. The Hosmer-Lemeshow test was used to verify goodness of fit of the model to the data. To determine whether a potential association with AH was independent of severity of the disease, we stratified patients into homogenous tertiles according to the model for end-stage liver disease (MELD) and used the Cochran-Mantel-Haenszel test. We explored the association of 25(OH)D with
346
mortality and complications of portal hypertension (ascites, encephalopathy, HRS, SBP, bleeding) in AC patients (n = 254) at one year. Follow-up time was defined as the period from the first day of medical therapy to the last followup visit. Patient survival, expressed in percentages, was estimated by the Kaplan-Meier method. The association between 25(OH)D levels and mortality/complications of portal hypertension was assessed by the log-rank test. With respect to the association with mortality, univariate, and multivariable (after adjustment for age, sex, BMI, seasonal timing of sample collection, and for the MELD) Cox proportional hazards models were performed. Data from in vitro and in vivo experiments are expressed as the median with the (minimum-maximum) values. All reported p-values are two-tailed. A p-value <0.05 was considered statistically significant. Statistical analyses were performed using SPSS version 19.0 software.
Results Association between 25(OH)D levels and clinical, biochemical, and histological parameters of ALD Healthy controls had a 25(OH)D median concentration of 19 ng/ ml, which was significantly higher than that of ALD patients (9.6 ng/ml) in univariate analysis (p = 4.75 1024, Table 1) and after adjustment for age, sex, BMI, and seasonal timing of sample collection (p = 1.45 108). After matching patients and healthy controls for age (±5 years) and sex (Supplementary Table 4) the difference in the 25(OH) median concentration remained highly significant in univariate (p = 3.46 1011) and multivariable analyses (p = 6.43 108). Clinical, biochemical and histological characteristics according to 25(OH)D levels (<10 ng/ml vs. P10 ng/ml) are summarized in Table 2. In addition, clinical, biological, and histological characteristics of cirrhotic and non-cirrhotic patients are reported in Supplementary Table 2. No association was observed between age, sex, BMI, and heavy alcohol consumption or 25(OH)D levels. Similarly, no significant association of 25(OH)D with ALT was seen, in contrast with AST (p = 1.00 103). Furthermore, patients with 25(OH)D below 10 ng/ml had a higher HVPG (p = 5.80 106). In addition, low 25(OH)D levels were strongly associated with severity of liver disease assessed by MELD (p = 2.50 104) and Child-Pugh scores (p = 8.50 107). Finally, a low vitamin D concentration was significantly associated with more severe steatosis (p = 1.00 103), fibrosis (p = 0.016), and AH (p = 2.14 104). After stratification of patients into tertiles according to the MELD score (<9; 9–16.2; >16.2), the association with AH remained statistically significant (p = 0.035). Conversely, we reported no association between vitamin D status and the Maddrey discriminant function (mDF) score in AH patients (p = 0.093), nor with the occurrence of HCC (p = 1.000). Remarkably, seasonal timing of sample collection for 25(OH)D measurement during summer and autumn was linked to higher levels compared to winter and spring months (p = 0.033). Association between 25(OH)D levels and cirrhosis and its related complications Low 25(OH)D concentrations were significantly associated with cirrhosis (odds ratio [OR] = 2.13, 95% confidence interval [CI] = 1.18–3.84, p = 0.013, Table 3) independently of other classical risk factors (age, sex, BMI, and heavy alcohol consumption) and seasonal timing of sample collection. In addition, we observed a significant link between cirrhosis and obesity (OR = 1.88, 95% CI = 1.04–3.40, p = 0.036, Table 3) and the
Journal of Hepatology 2013 vol. 59 j 344–350
JOURNAL OF HEPATOLOGY Table 2. Association of serum 25(OH)D levels with clinical, biological, and histological variables in ALD patients.
<10 ng/ml (n = 171) 38.0 53.4 ± 9.0 56.3 39.2 33.0 (25-55) 60 (43-107) 143.0 (59.3-323) 3.1 (1.2-9.3) 1.46 (1.12-1.78) 16 (12-21) 16.6/28.6/54.8 14.4 (8.8-21.0) 36.0 (23.8-59.0) 78.0 29.4 20 (9-50) 1.2/2.9/2.3/9.4/84.2 2.3
Female sex, % Age, yr BMI >25 kg/m2, % Blood sampling during summer/autumn, % ALT, IU/L AST, IU/L GGT, IU/L Bilirubin, mg/dl INR HVPG, mmHg Child-Pugh score A/B/C, % MELD score mDF Alcohol >80 g/day, % AH, % Steatosis, % Fibrosis stage 0/1/2/3/4, %* HCC, %
25(OH)D levels ≥10 ng/ml (n = 153) 32.0 55.4 ± 9.6 61.6 51.0 35 (24-49) 48 (34-78) 146.5 (72.0-297.0) 1.3 (0.7-2.7) 1.27 (1.08-1.54) 12 (6-17) 36.4/41.8/21.8 10.0 (6.4-15.4) 27.0 (7.8-45.8) 73.5 10.3 10 (5-35) 2/5.6/12.5/8.3/71.6 2.6
p value
0.260 0.057 0.339 0.033 0.874 1.00 x 10-3 0.816 1.42 x 10-7 3.20 x 10-4 5.80 x 10-6 8.50 x 10-7 2.50 x 10-4 0.071 0.356 7.01 x 10-5 1.00 x 10-3 0.016 1.000
25(OH)D, 25 hydroxyvitamin D; BMI, body mass index; ALT, alanine aminotransferase; AST, aspartate aminotransferase; GGT, gamma-glutamyl transferase; HCC, hepatocellular carcinoma; HVPG, hepatic venous pressure gradient; MELD, model for end-stage liver disease; AH, alcoholic hepatitis; mDF, Maddrey discriminant function test. Variables with normal distribution are presented as mean ± standard deviation. Skewed variables are presented as median (interquartile range). ⁄ Available in 97% of cases, including all non-cirrhotic patients.
Table 3. Univariate and multivariable analysis of risk of cirrhosis.
Variables Age Sex (female) BMI >25 kg/m2 Alcohol >80 g/day PNPLA3 rs738409 25(OH)D <10 ng/ml*
OR 1.02 1.14 2.15 1.35 2.24 1.73
Univariate 95% CI 0.99-1.05 0.65-2.00 1.45-3.57 0.74-2.46 1.31-3.83 1.01-2.95
p value 0.097 0.645 6.00 x 10-3 0.329 3.00 x 10-3 0.044
OR 1.03 1.39 1.88 1.71 2.42 2.13
Multivariable* 95% CI 0.99-1.06 0.75-2.56 1.04-3.40 0.89-3.29 1.36-4.30 1.18-3.84
p value 0.070 0.270 0.036 0.109 3.00 x 10-3 0.013
OR, odds ratio; CI, confidence interval; 25(OH)D, 25 hydroxyvitamin D; BMI, body mass index. ⁄ Adjusted for seasonal timing of sample collection.
rs738409 variant in the PNPLA3 gene (OR = 2.42, 95% CI = 1.36– 4.30, p = 3.00 103, Table 3). Allele frequencies for the PNPLA3 SNP were in Hardy-Weinberg equilibrium, indicating that genotype and allele frequencies were in equilibrium and suggesting the absence of a problem in genotyping (data not shown) [22]. The Hosmer–Lemeshow test was not significant (p = 0.644), suggesting good calibration of the model. Furthermore, a 25(OH)D concentration below 10 ng/ml was significantly associated with complications of portal hypertension at one year after adjustment for seasonal timing of sample collection and disease severity (MELD score in tertiles), including the presence of ascites (p = 1.22 104), encephalopathy (p = 5.22 103) and HRS (p = 0.025) (Supplementary Table 3). However, although SBP and bleeding were more prevalent in the low vitamin D concentration group, this difference was not statistically significant (p = 0.451 and p = 0.449, Supplementary Table 3). The proportion of patients lost to follow-up was nearly 50% for most events studied, but was not statistically different
between patients with 25(OH)D <10 and P10 ng/ml (data not shown). Association between 25(OH)D levels and mortality in AC patients The median follow-up was 4.9 months. After 12 months, 12.6% died (91% of this mortality was liver-related) and 11.4% underwent liver transplantation (Supplementary Table 2). During the same period, the proportion of patients lost to follow-up was 52% at 12 months. Nevertheless, this proportion was not statistically different in patients with or without severe vitamin D deficiency (48.3% vs. 57.9%, p = 0.129). A 25(OH)D concentration below 10 ng/ml was significantly linked to a lower overall survival rate at one year (hazards ratio [HR] = 5.95 95% CI = 2.09–16.96, p = 8.56 104, Fig. 1 and Table 4). In multivariable analysis, only the MELD score and severe 25(OH)D deficiency remained associated with one-year mortality in AC patients (HR = 4.33, 95% CI = 1.47–12.78, p = 7.94 103, Table 4).
Journal of Hepatology 2013 vol. 59 j 344–350
347
Research Article **
8000
0.9 TNFα (pg/ml)
p = 8.56 x 10-4
0.8 0.7
25(OH)D <10 ng/ml
*
HS AC
6000 4000 §§
§
2000
0.6
Vit D ≥10 112 Vit D <10 142
2 74 78
4 64 68
6 8 10 Follow-up (months) 58 63
51 55
45 52
12
14
43 45
Fig. 1. Kaplan Meier curves illustrating the probability of survival of alcoholic cirrhotic (AC) patients according to 25-hydroxyvitamin D [25(OH)D] cut-offs (<10 vs. P10 ng/ml). A total of 254 AC patients were followed for one year. The table below indicates the number of patients remaining at risk of mortality at each time point.
Influence of 1,25(OH)2D on TNFa production by stimulated PBMCs In PBMCs isolated from AC patients (n = 8; 25(OH)D = 10 [7– 60.2] ng/ml) and from healthy subjects (HS; n = 8; 25(OH)D = 27 [22–34] ng/ml), ConA stimulation induced a significant increase in TNFa production (HS, TNFa without ConA: 191.5 [0–614.8] pg/ml vs. TNFa after ConA: 5183.4 [1735.2–9859.8] pg/ ml, p = 3.00 103; AC, TNFa without ConA: not detectable vs. TNFa after ConA: 7487.7 [2377.5–12,000] pg/ml, p = 3.31 104), while 1 h pretreatment with 107 M 1,25(OH)2D significantly decreased TNFa production both in HS and in AC patients (HS, TNFa after 1,25(OH)2D + ConA: 1556.0 [873.2–3781.1] pg/ml vs. TNFa after ConA, p = 0.012; AC TNFa after 1,25(OH)2D + ConA: 1693.7 [1106.1–4314.7] pg/ml vs. TNFa after ConA, p = 3.00 103; Fig. 2). Influence of 1,25(OH)2D supplementation in a murine experimental model of ALD Serum levels of ALT and the liver/mouse weight ratio were both increased in the EtOH group compared to the Ctrl group, as expected from the experimental model (ALT: EtOH 50 [10– 100] IU/L vs. Ctrl 10 [10–32] IU/L, p = 1.40 104; liver/mouse weight ratio: EtOH 7.0 [6.2–8.1] vs. Ctrl 5.6 [4.8–6.3], p = 9.00 106). However, vitamin D supplementation had no significant impact on these variables (ALT: EtOH + 1,25(OH)2D 56 [36–132] pg/ml; liver/mouse weight ratio: EtOH + 1,25(OH)2D 7.1 [6.2–10.8], Fig. 3A and B). TNFa expression was induced in livers of alcohol-fed mice compared to the Ctrl group (TNFa expression in EtOH mice: 2.4
n.d.
0 ConA (2.5 µg/m l) 1,25(OH)2D (10-7M)
-
n.d.
+
+ -
+ +
Fig. 2. Effect of 1,25(OH)2D on TNFa production by stimulated PBMCs. PBMCs, isolated from patients with alcoholic cirrhosis (AC; n = 8) and healthy subjects (HS; n = 8), were pretreated with 107 M 1,25-dihydroxyvitamin D [1,25(OH)2D] or with vehicle control before stimulation with 2.5 lg/ml ConA. TNFa production was assayed in culture supernatants by ELISA. Bars represent median with interquartile range (duplicate cultures for each point). No stimulation vs. ConA stimulation, ⁄p = 3.00 103 in HS; ⁄⁄p = 3.31 104 in AC; ConA vs. 1,25(OH)2D + ConA; §p = 0.012 in HS; §§p = 3.00 103 in AC; n.d., not detectable.
A
Control EtOH EtOH + 1,25(OH)2D 80 *
§
60 ALT (IU/L)
0
40 20
B
Control EtOH EtOH + 1,25(OH)2D 10
Liver/mouse weight ratio
Survival probability
10,000
25(OH)D ≥10 ng/ml
1.0
8
*
§
6 4 2
0 Fig. 3. Effect of 1,25(OH)2D oral supplementation in a murine experimental model of ALD. (A) Plasma alanine aminotransferase (ALT) and (B) liver/mouse weight ratio assessment in mice fed the Lieber-DeCarli diet for 6 weeks. Eighteen mice from the ethanol group were orally administered 50 ng of 1,25(OH)2D three times a week for 5 weeks (EtOH + 1,25(OH)2D), while pair-fed (Ctrl; n = 7) and ethanol-fed (EtOH; n = 22) groups received PBS as control vehicle. Values are expressed as median and interquartile range. ALT, ⁄p = 1.40 104 EtOH vs. Ctrl; § p = 1.20 105 EtOH + 1,25(OH)2D vs. Ctrl; liver/mouse weight ratio, ⁄ p = 9.00 106 EtOH vs. Ctrl; §p = 2.40 105 EtOH + 1,25(OH)2D vs. Ctrl.
[1.0–7.1] times the expression in Ctrl mice, p = 1.00 103), while oral administration of 1,25(OH)2D significantly decreased expression of TNFa in the livers of EtOH-fed mice by 35% (p = 0.04, Fig. 4).
Table 4. Univariate and multivariable analysis of risk of mortality at one year.
Variables Age Sex (female) BMI >25 kg/m2 MELD score 25(OH)D <10 ng/ml*
HR 1.00 1.31 1.46 1.09 5.95
Univariate 95% CI 0.97-1.04 0.65-2.66 0.69-3.08 1.06-1.12 2.09-16.96
p value 0.837 0.451 0.323 1.62 x 10-9 8.56 x 10-4
HR 1.02 1.31 1.07 1.08 4.33
Multivariable* 95% CI 0.98-1.07 0.63-2.73 0.48-2.42 1.04-1.12 1.47-12.78
HR, hazards ratio; CI, confidence interval; 25(OH)D, 25 hydroxyvitamin D; BMI, body mass index; MELD, model for end-stage liver disease. ⁄ Adjusted for seasonal timing of sample collection.
348
Journal of Hepatology 2013 vol. 59 j 344–350
p value 0.295 0.471 0.864 1.15 x 10-6 7.94 x 10-3
JOURNAL OF HEPATOLOGY Control EtOH EtOH + 1,25(OH)2D Hepatic TNFα mRNA (fold increase)
4 * §
3
**
2 1 0
Fig. 4. Effect of 1,25(OH)2D oral supplementation on hepatic expression of TNFa after 6 weeks of the Lieber-DeCarli diet. Gene transcripts were analyzed by RT-PCR and results were normalized to HPRT mRNA levels using the 2DDCT method. Data are expressed as fold increase in normalized TNFa expression by livers of the pair-fed group (Ctrl; n = 7). Bars represent median with interquartile range. ⁄p = 1.00 103 EtOH vs. Ctrl; ⁄⁄p = 0.021 ethanol-fed mice orally supplemented with 1,25(OH)2D (EtOH + 1,25(OH)2D; n = 18) vs. Ctrl; §p = 0.04 ethanolfed mice (EtOH; n = 22) vs. EtOH + 1,25(OH)2D.
Discussion This study demonstrates for the first time that low serum levels of 25(OH)D are associated with increased liver damage (AST levels, steatosis, AH, and cirrhosis), and especially with more advanced liver disease (higher HVPG levels and MELD/Child-Pugh scores) in ALD patients. Furthermore, a low threshold (<10 ng/ml) was significantly linked to increased mortality. Low serum levels of vitamin D have long been observed in chronic liver diseases, especially at the cirrhosis stage [9]. They were also shown to be associated with liver damage, especially fibrosis, in NAFLD and CHC [11–13]. In line with these observations, we reported that low 25(OH)D concentrations were associated with cirrhosis and that the link remained significant even after adjustment for other classical risk factors (age, sex, alcohol consumption, and BMI), as was the PNPLA3 (rs738409 C>G) variant [4–6]. Moreover, like previous authors, we observed that the vitamin D concentration was closely related to liver disease severity as assessed by Child-Pugh [23], and MELD scores in AC patients. We also reported a novel link to AH, portal hypertension, and most of its related complications at one year. Interestingly, the association with AH was independent of severity of the disease. In line with the previous considerations, the link between higher mortality and severe vitamin D deficiency was therefore expected. Thus, a low 25(OH)D concentration remained associated with mortality independently of the MELD score and was the leading risk factor associated with short-term mortality in liver cirrhosis, suggesting that severe 25(OH)D deficiency is a marker of AC severity/prognosis. Generally speaking, the proportion of patients lost to follow-up was substantial (nearly 50% at one year for most events studied). This limitation prevented an accurate analysis over a longer period of time. Nonetheless, the percentage of lost to follow-up patients was not statistically different between patients with severe 25(OH)D deficiency and those without. Overall, although these associations are interesting, further studies using a prospective design are required to confirm these findings. In addition, severe vitamin D deficiency was not associated with HCC; however, this analysis again requires prospective cohorts of AC followed for a much longer time to reach a firm conclusion.
Sources of vitamin D include increased production through sunlight and, to a much lower extent, diet and dietary supplements. Nutritional status was not assessed in this study, which may be considered another limitation. Nevertheless, adequate assessment of malnutrition is particularly challenging in ALD [24]. Moreover, the effect of dietary intake on 25(OH)D status is relatively limited, since few foods contain sizable amounts of vitamin D. Conversely, the vast majority of 25(OH)D is derived from the skin through sunlight exposure [9]. Thus, based on observations in NAFLD and CHC, we also report that seasonal timing of the sampling had a significant influence on vitamin D status in ALD patients [10,12]. Overall, these association data using a cross-sectional design do not permit establishing a causal relationship between vitamin D severe deficiency and ALD severity/mortality. However, growing evidence suggests that 25(OH)D might play a biological role in various liver diseases [10,11,13]. The increased levels of TNFa found in ALD patients are associated with development of complications and with the severity and mortality of cirrhosis [7]. Indeed, a leading theory in ALD pathogenesis suggests the deleterious consequences of a pro-inflammatory cascade, favored by ethanol and followed by a dysregulated immune response to stimulation [25]. Experimental studies have shown that 1,25(OH)2D can suppress the release of TNFa both in vivo [26,27] and in vitro [28]. Interestingly, our in vitro experiments confirmed the inhibitory effect of 1,25(OH)2D on TNFa production and suggest that this inhibitory mechanism is not altered by hepatic disease. Furthermore, in our in vivo experiments, 1,25(OH)2D oral supplementation reduced hepatic expression of TNFa induced by an ethanol diet. Although correction of vitamin D status did not modify transaminase serum levels, these results, in line with in vitro human data on PBMCs, suggest that it may downplay the deleterious inflammatory cascade that takes place in the liver following excessive alcohol consumption, by reactivating the functional inhibitory pathway. This hypothesis is further supported by a recently published study demonstrating that vitamin D deficiency exacerbates the development of NAFLD in rats through increased expression of inflammatory genes in the liver [29]. In conclusion, this study reports that 25(OH)D levels are significantly associated with ALD severity, portal hypertension, its related complications and mortality at one year. In addition, our data suggest that 1,25(OH)2D could improve the deleterious pro-inflammatory cytokine profile of AC patients. Although our results must be replicated in prospective cohorts to address the true impact of 25(OH)D severe deficiency on the ALD phenotype, vitamin D may well represent both a biomarker and a therapeutic target worth exploring in ALD. Financial support This work was supported by the Fonds Erasme for Medical Research and the FNRS (Belgian National Fund of Scientific Research).
Conflict of interest The authors who have taken part in this study declared that they do not have anything to disclose regarding funding or conflict of interest with respect to this manuscript.
Journal of Hepatology 2013 vol. 59 j 344–350
349
Research Article Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jhep.2013. 03.024.
[14]
[15] [16]
References [1] Rehm J, Mathers C, Popova S, Thavorncharoensap M, Teerawattananon Y, Patra J. Global burden of disease and injury and economic cost attributable to alcohol use and alcohol-use disorders. Lancet 2009;373:2223–2233. [2] European, Association for the Study of the Liver. EASL clinical practical guidelines: management of alcoholic liver disease. J Hepatol 2012;57:399–420. [3] Rehm J, Taylor B, Mohapatra S, Irving H, Baliunas D, Patra J, et al. Alcohol as a risk factor for liver cirrhosis: a systematic review and meta-analysis. Drug Alcohol Rev 2010;29:437–445. [4] Naveau S, Giraud V, Borotto E, Aubert A, Capron F, Chaput JC. Excess weight risk factor for alcoholic liver disease. Hepatology 1997;25:108–111. [5] Stickel F, Buch S, Lau K, Zu Schwabedissen HM, Berg T, Ridinger M, et al. Genetic variation in the PNPLA3 gene is associated with alcoholic liver injury in caucasians. Hepatology 2011;53:86–95. [6] Trépo E, Gustot T, Degre D, Lemmers A, Verset L, Demetter P, et al. Common polymorphism in the PNPLA3/adiponutrin gene confers higher risk of cirrhosis and liver damage in alcoholic liver disease. J Hepatol 2011;55:906–912. [7] Gao B, Bataller R. Alcoholic liver disease: pathogenesis and new therapeutic targets. Gastroenterology 2011;141:1572–1585. [8] Holick MF. Vitamin D deficiency. N Engl J Med 2007;357:266–281. [9] Kitson MT, Roberts SK. D-livering the message: the importance of vitamin D status in chronic liver disease. J Hepatol 2012;57:897–909. [10] Targher G, Bertolini L, Scala L, Cigolini M, Zenari L, Falezza G, et al. Associations between serum 25-hydroxyvitamin D3 concentrations and liver histology in patients with non-alcoholic fatty liver disease. Nutr Metab Cardiovasc Dis 2007;17:517–524. [11] Petta S, Camma C, Scazzone C, Tripodo C, Di Marco V, Bono A, et al. Low vitamin D serum level is related to severe fibrosis and low responsiveness to interferon-based therapy in genotype 1 chronic hepatitis C. Hepatology 2010;51:1158–1167. [12] Bitetto D, Fattovich G, Fabris C, Ceriani E, Falleti E, Fornasiere E, et al. Complementary role of vitamin D deficiency and the interleukin-28B rs12979860 C/T polymorphism in predicting antiviral response in chronic hepatitis C. Hepatology 2011;53:1118–1126. [13] Lange CM, Bojunga J, Ramos-Lopez E, von Wagner M, Hassler A, Vermehren J, et al. Vitamin D deficiency and a CYP27B1-1260 promoter polymorphism are
350
[17]
[18]
[19]
[20]
[21]
[22] [23]
[24] [25] [26]
[27] [28]
[29]
associated with chronic hepatitis C and poor response to interferon-alfa based therapy. J Hepatol 2011;54:887–893. Faria SC, Ganesan K, Mwangi I, Shiehmorteza M, Viamonte B, Mazhar S, et al. MR imaging of liver fibrosis: current state of the art. Radiographics 2009;29:1615–1635. Lucey MR, Mathurin P, Morgan TR. Alcoholic hepatitis. N Engl J Med 2009;360:2758–2769. European, Association for the Study of the Liver, European, Organisation for Research and Treatment of Cancer. EASL-EORTC clinical practice guidelines: management of hepatocellular carcinoma. J Hepatol 2012;56: 908–943. European, Association for the Study of the Liver. EASL clinical practice guidelines on the management of ascites, spontaneous bacterial peritonitis, and hepatorenal syndrome in cirrhosis. J Hepatol 2010;53: 397–417. Brunt EM, Janney CG, Di Bisceglie AM, Neuschwander-Tetri BA, Bacon BR. Nonalcoholic steatohepatitis: a proposal for grading and staging the histological lesions. Am J Gastroenterol 1999;94:2467–2474. Bosch J, Abraldes JG, Berzigotti A, Garcia-Pagan JC. The clinical use of HVPG measurements in chronic liver disease. Nat Rev Gastroenterol Hepatol 2009;6:573–582. McMullen MR, Pritchard MT, Wang Q, Millward CA, Croniger CM, Nagy LE. Early growth response-1 transcription factor is essential for ethanol-induced fatty liver injury in mice. Gastroenterology 2005;128:2066–2076. Bischoff-Ferrari HA, Giovannucci E, Willett WC, Dietrich T, Dawson-Hughes B. Estimation of optimal serum concentrations of 25-hydroxyvitamin D for multiple health outcomes. Am J Clin Nutr 2006;84:18–28. Balding DJ. A tutorial on statistical methods for population association studies. Nat Rev Genet 2006;7:781–791. Malham M, Jorgensen SP, Ott P, Agnholt J, Vilstrup H, Borre M, et al. Vitamin D deficiency in cirrhosis relates to liver dysfunction rather than aetiology. World J Gastroenterol 2011;17:922–925. McClain CJ, Barve SS, Barve A, Marsano L. Alcoholic liver disease and malnutrition. Alcohol Clin Exp Res 2011;35:815–820. Tsukamoto H, Lu SC. Current concepts in the pathogenesis of alcoholic liver injury. FASEB J 2001;15:1335–1349. Zhu Y, Mahon BD, Froicu M, Cantorna MT. Calcium and 1 alpha,25dihydroxyvitamin D3 target the TNF-alpha pathway to suppress experimental inflammatory bowel disease. Eur J Immunol 2005;35:217–224. Lemire JM. Immunomodulatory role of 1,25-dihydroxyvitamin D3. J Cell Biochem 1992;49:26–31. Muller K, Haahr PM, Diamant M, Rieneck K, Kharazmi A, Bendtzen K. 1,25Dihydroxyvitamin D3 inhibits cytokine production by human blood monocytes at the post-transcriptional level. Cytokine 1992;4:506–512. Roth CL, Elfers CT, Figlewicz DP, Melhorn SJ, Morton GJ, Hoofnagle A, et al. Vitamin D deficiency in obese rats exacerbates nonalcoholic fatty liver disease and increases hepatic resistin and Toll-like receptor activation. Hepatology 2012;55:1103–1111.
Journal of Hepatology 2013 vol. 59 j 344–350
Marked 25-hydroxyvitamin D deficiency is associated with poor prognosis in patients with alcoholic liver disease Eric Trépo1,2, , Romy Ouziel1,2, , Pierre Pradat3,4,5, Yukihide Momozawa6, Eric Quertinmont2, Christine Gervy7, Thierry Gustot1,2, Delphine Degré1,2, Vincent Vercruysse2, Pierre Deltenre8, Laurine Verset9, Beatrice Gulbis7, Denis Franchimont1,2, Jacques Devière1,2, Arnaud Lemmers1,2, Christophe Moreno1,2,⇑ 1
Department of Gastroenterology, Hepatopancreatology and Digestive Oncology, Erasme Hospital, Université Libre de Bruxelles, Brussels, Belgium; 2Laboratory of Experimental Gastroenterology, Université Libre de Bruxelles, Brussels, Belgium; 3Hospices Civils de Lyon, Hôpital de la Croix-Rousse, Lyon, France; 4INSERM, U1052, Lyon, France; 5Université Lyon 1, Lyon, France; 6Unit of Animal Genomics, Groupe Interdisciplinaire de Génoprotéomique Appliquée (GIGA-R) and Faculty of Veterinary Medicine, University of Liège (B34), Liège, Belgium; 7 Department of Clinical Chemistry, Erasme Hospital, Université Libre de Bruxelles, Brussels, Belgium; 8Service d’Hépato-Gastroentérologie, Hôpital de Jolimont, Haine-Saint-Paul, Belgium; 9Department of Pathology, Erasme Hospital, Université Libre de Bruxelles, Brussels, Belgium
Background & Aims: Vitamin D deficiency has been frequently reported in advanced liver disease. However, its influence on alcoholic liver disease (ALD) has been poorly elucidated. We investigated the association of vitamin D with clinical, biological, and histological parameters and survival in ALD patients. Furthermore, we explored the effect of vitamin D treatment on ALD patient peripheral blood mononuclear cells (PBMCs), and in a murine experimental model of ALD. Methods: Serum levels of 25-hydroxyvitamin D [25(OH)D] were determined in 324 Caucasian ALD patients and 201 healthy controls. In vitro experiments on vitamin D pre-treated PBMCs evaluated TNFa production by ELISA in culture supernatants. Mice were submitted to an ethanol-fed diet and some of them were orally supplemented three times per week with 1,25(OH)2D.
Keywords: Vitamin D; Genetic factors; Alcoholic liver disease; Inflammation, TNFa.
Results: Severe deficiency in 25(OH)D (<10 ng/ml) was significantly associated with higher aspartate aminotransferase levels (p = 1.00 103), increased hepatic venous pressure gradient (p = 5.80 106), MELD (p = 2.50 104), and Child-Pugh scores (p = 8.50 107). Furthermore, in multivariable analysis, a low 25(OH)D concentration was associated with cirrhosis (OR = 2.13, 95% CI = 1.18–3.84, p = 0.013) and mortality (HR = 4.33, 95% CI = 1.47–12.78, p = 7.94 103) at one year. In addition, in vitro, 1,25(OH)2D pretreatment decreased TNFa production by stimulated PBMCs of ALD patients (p = 3.00 103), while in vivo, it decreased hepatic TNFa expression in ethanolfed mice (p = 0.04). Conclusions: Low 25(OH)D levels are associated with increased liver damage and mortality in ALD. Our results suggest that vitamin D might be both a biomarker of severity and a potential therapeutic target in ALD. Ó 2013 European Association for the Study of the Liver. Published by Elsevier B.V. All rights reserved.
Received 6 July 2012; received in revised form 14 March 2013; accepted 15 March 2013; available online 1 April 2013 ⇑ Corresponding author. Address: Erasme Hospital, Department of Gastroenterology, Hepatopancreatology and Digestive Oncology. Université Libre de Bruxelles. Route de Lennik, 808 1070 Brussels, Belgium. Tel.: +32 2 555 37 12; fax: +32 2 555 46 97. E-mail address: [email protected] (C. Moreno). These authors contributed equally to this work. Abbreviations: HCC, hepatocellular carcinoma; ALD, alcoholic liver disease; BMI, body mass index; AC, alcoholic cirrhosis; NAFLD, non-alcoholic fatty liver disease; CHC, chronic hepatitis C; 25(OH)D, 25-hydroxyvitamin D; 1,25(OH)2D, 1,25-dihydroxyvitamin D; GWAS, genome-wide association studies; PBMCs, peripheral blood mononuclear cells; ALT, alanine aminotransferase; AST, aspartate aminotransferase; AH, alcoholic hepatitis; GGT, gamma-glutamyl transferase; SBP, spontaneous bacterial peritonitis; HRS, hepatorenal syndrome; HVPG, hepatic venous pressure gradient; SNP, single nucleotide polymorphism; ConA, concanavalin A; ELISA, enzyme-linked immunosorbent assay; TNFa, tumor necrosis factor-a; RT-PCR, reverse transcriptase polymerase chain reaction; HPRT, hypoxanthine-guanine phosphoribosyltransferase; EtOH, ethanol; Ctrl, control; MELD, model for end-stage liver disease; OR, odds ratio; CI, confidence interval; HR, hazards ratio; HS, healthy subjects.
Introduction Nearly 4% of worldwide mortality is related to excessive alcohol consumption [1]. More specifically, it is a leading cause of cirrhosis and hepatocellular carcinoma (HCC), which may ultimately requires liver transplantation [2]. Alcoholic liver disease (ALD) and its complications increase in proportion to daily ethanol consumption [3]. Moreover, clinical factors (older age, female sex, body mass index [BMI]) have been associated with a higher prevalence of alcoholic cirrhosis (AC) [4]. In addition, a genetic variant in the PNPLA3 (rs738409 C>G) gene was recently linked to AC in Caucasians [5,6]. Nonetheless, new diagnostic biomarkers and therapeutic targets are urgently needed when confronting alcohol-related liver injury [7].
Journal of Hepatology 2013 vol. 59 j 344–350
JOURNAL OF HEPATOLOGY Vitamin D deficiency has been associated with various skeletal and non-skeletal chronic illnesses [8]. In particular, it has also long been reported in various chronic liver diseases, especially at the cirrhosis stage [9]. More recent studies observed that low vitamin D serum levels were associated with increased histological liver damage in non-alcoholic fatty liver disease (NAFLD) [10] and in chronic hepatitis C (CHC) [11–13]. Whatever its origin (via skin production, diet or dietary supplements), vitamin D must undergo two hydroxylations in order to become biologically active: first in the liver, generating 25-hydroxyvitamin D [25(OH)D], the most relevant indicator of a patient’s vitamin D status; and then in the kidney, as 1,25-dihydroxyvitamin D [1,25(OH)2D]. In addition to its role in bone metabolism and calcium homeostasis, 1,25(OH)2D, the active form of vitamin D, has antitumoral activity, antifibrogenic effects, and anti-inflammatory properties [9]. The aim of the present study was to investigate the association of 25(OH)D deficiency with histological damage, portal hypertension and its related complications, liver function, and mortality in ALD. In addition, since the immunomodulating role of vitamin D has been established, we studied the influence of 1,25(OH)2D treatment upon inflammation in ALD patient peripheral blood mononuclear cells (PBMCs) in vitro, and we evaluated the effects of 1,25(OH)2D oral supplementation in a murine experimental model of ALD in vivo.
For each patient, the date of inclusion was the date of the first liver biopsy or, in its absence, the first blood sampling at the time of admission. Gender, age, BMI, past history of diabetes mellitus, daily alcohol intake, presence of complications of portal hypertension (ascites, hepatic encephalopathy HRS, SBP or bleeding), serum 25(OH)D bilirubin, albumin, prothrombin, gammaglutamyl transferase, ALT, and AST levels were recorded at inclusion. All patients were followed up after discharge from the liver unit. Thus, they were scheduled for new evaluations at a minimum of 6-month intervals, or sooner in case of more severe disease. In addition to conventional monitoring, HCC screening included liver ultrasonography and serum AFP level assessment. Follow-up ended at the date of death or liver transplantation (the patient was considered alive), or at the last recorded visit or information gathering. This date was set as the final time limit for upgrading the patient file using our hospital computerized database. Patients transplanted or lost to followup were included in the analysis and censored at the date of transplantation or last recorded information. In addition, 201 healthy Caucasian social workers from the same geographic area (Brussels area) as ALD patients were included as controls. They had neither clinical nor biochemical evidence (routine blood screening) of liver disease, nor evidence of other major pathological conditions. This group served as a reference for vitamin D levels. None of the patients or healthy controls included in our study used vitamin D supplements or medication known to affect the serum vitamin D concentration. Written informed consent was obtained from all patients and this study was approved by the local ethics committee. Patient and healthy control characteristics are summarized in Table 1.
Biological variables assessment Biological parameters were measured concomitantly with the liver biopsy procedure or at the time of admission. Quantitative assessment of serum total 25(OH)D levels (25-hydroxyvitamin D2 and D3) was performed by direct competitive chemiluminescence immunoassay using commercially available kits from Diasorin on a LiaisonÒ (Diasorin, Stillwater, MN, USA).
Patients and methods Patients A total of 324 ALD patients were recruited prospectively between January 1, 2002 and May 31, 2011 at Erasme Hospital, Brussels, Belgium. According to the ALD definition, these unrelated Caucasian patients had a history of excessive alcohol intake of >30 g/day (76% drank more than 80 g/day), together with increased alanine aminotransferase (ALT) and aspartate aminotransferase (AST) or suspected AC [2]. The diagnosis of cirrhosis was based on liver biopsy and/ or an unequivocal clinico-biochemical profile, a highly suggestive endoscopic exam (e.g., esophageal varices with or without a previous episode of bleeding) or typical imaging findings in agreement with current guidelines [14]. Seventyeight percent of these ALD patients had cirrhosis. The diagnosis of alcoholic hepatitis (AH) was based on both compatible histological signs and a history of recent jaundice onset, with serum bilirubin >5 mg/dl [15]. The presence of HCC was based on computed tomodensitometry and/or magnetic-resonance imaging and/or guided liver biopsy [16]. The diagnosis of spontaneous bacterial peritonitis (SBP) was based on an ascitic fluid neutrophil count P250 cells/ mm3, and that of hepatorenal syndrome (HRS) was in agreement with current guidelines [17]. Patients with other causes of liver disease, including viral or autoimmune hepatitis, hemochromatosis or co-infection with human immunodeficiency virus, were excluded.
Histological assessment Ninety-seven percent of all ALD patients underwent liver biopsy. The percentage of steatosis was determined among the total number of hepatocytes, and fibrosis was assessed according to the Brunt score [18], using Sirius Red staining. Diagnosis of AH was established according to current guidelines [15].
Hepatic venous pressure gradient assessment (HVPG) Using a straight catheter, measurement of HVPG was performed concomitantly with transjugular biopsy in 91% of ALD patients for whom cirrhosis was suspected. This indication was based on the excellent correlation between HVPG and portal hypertension, as was the high predictive value of cirrhosis decompensation [19]. HVPG was calculated as the difference between wedge hepatic venous pressure and free hepatic venous pressure [19]. Each measurement was repeated three times.
Table 1. Characteristics of patients (n = 324) and healthy controls (n = 201).
Female sex, % Age, yr BMI, kg/m2 BMI >25 kg/m2, % 25(OH)D, ng/ml ALT, IU/L AST, IU/L GGT, IU/L Cirrhosis, % AH, %
ALD patients 35.2 54.3 ± 9.4 26.0 ± 5.3 54.3 9.6 (5.0-16.2) 34 (24-49) 55 (38-92) 143.5 (65.3-308) 78.4 20.8
Controls 50.2 43.8 ± 9.7 25.2 ± 5.4 45.5 19.0 (13.5-25.0) 15 (11-21) 24 (19-29) 21.3 (14.4-32.4) -
p value 1.00 x 10-3 1.13 x 10-30 0.084 0.057 4.75 x 10-24 8.51 x 10-45 1.34 x 10-53 1.61 x 10-54 -
BMI, body mass index; 25(OH)D, 25 hydroxyvitamin D; ALT, alanine aminotransferase; AST, aspartate aminotransferase; GGT, gamma-glutamyl transferase; AH, alcoholic hepatitis. Variables with normal distribution are presented as mean ± standard deviation. Skewed variables are presented as median (interquartile range).
Journal of Hepatology 2013 vol. 59 j 344–350
345
Research Article Genotyping DNA was isolated using the phenol chloroform method from a whole blood sample collected concomitantly with liver biopsy. Twenty nanograms of DNA, TaqMan assays (Applied Biosystems, Foster City, CA, USA) and the LightCyclerÒ 480 system (Roche Diagnostics, Mannheim, Germany) were used to genotype PNPLA3 (rs738409, assay ID C_7241_10). Peripheral blood mononuclear cells (PBMCs) PBMCs from 8 healthy volunteers and 8 consecutive AC patients admitted to our institution were freshly isolated using a Ficoll gradient. Then, cells were seeded on 96-well culture plates for 1 h treatment with 107 M 1,25(OH)2D (Sigma Aldrich) or ethanol 0.1% as vehicle control. Thereafter, pre-treated PBMCs were stimulated with concanavalin A (ConA; 2.5 lg/ml) (Sigma Aldrich). Cell-free supernatants were collected 24 h later and stored at 20 °C until the enzymelinked immunosorbent assay (ELISA). Immunoassays An ELISA commercially available kit (Duoset, R&D systems) was used to quantify human tumor necrosis factor a (TNFa). In vivo experiment Eight-week old female C57BL/6 mice (Charles River, Brussels, Belgium) were submitted to the Lieber-DeCarli diet for 6 weeks following a previously published protocol [20]. Some of the mice from the ethanol-fed group (EtOH) were orally supplemented three times per week with 1,25(OH)2D (50 ng; Sigma Aldrich) (EtOH + 1,25(OH)2D; n = 18). Pair-fed (Ctrl; n = 7) and EtOH (n = 22) groups received PBS as vehicle control (protocol detailed in the supplementary section). At the end of the experiment, mice were killed and livers were excised, weighed and frozen at 80 °C, and serum levels of ALT were measured with a commercially available kit (Roche/Hitachi). To explore the potential impact on hepatomegaly (a marker of steatosis) in mice, we calculated a liver/mouse weight ratio as follows: liver/mouse weight ratio = (liver weight/mouse weight) 100. Quantitative reverse transcriptase polymerase chain reaction (RT-PCR) Frozen murine liver samples were first homogenized in lysis solution with a MagNalyser (Roche Diagnostics). Then, total RNA was extracted with the High Pure RNA tissue kit (Roche Diagnostics) according to the manufacturer’s protocol. Reverse transcription and quantitative PCR carried out on murine livers for TNFa and hypoxanthine-guanine phosphoribosyltransferase (HPRT) are detailed in the Supplementary data. The sequences of PCR probes and primers are available in Supplementary Table 1. Statistical analysis 25(OH)D levels are categorized throughout the manuscript using a cut-off of 10 ng/ml for ALD patients and 20 ng/ml for healthy controls. These thresholds were chosen because they were highly similar to the median in both populations (Table 1) and also reflected admitted clinical cut-offs for vitamin D deficiency (<20 ng/ml) and severe deficiency (<10 ng/ml) [21]. We tested the association between 25(OH)D concentrations and seasonal timing of sample collection. Therefore, we compared patients who underwent blood sampling during summer and autumn to those recruited during months with lower sun exposure (winter-spring). Categorical variables were studied using the two-sided Chi-square test, the Chi-square test for trend or the twosided Fisher’s exact test, whereas quantitative variables were assessed using analysis of variance or non-parametric Kruskall-Wallis or Mann-Whitney tests when appropriate. We conducted univariate and multivariable logistic regression to investigate whether 25(OH)D low levels were associated with cirrhosis after adjustment for other clinical (age, sex, BMI, heavy alcohol consumption, and seasonal timing of sample collection) [3,4] and genetic (rs738409 variant in PNPLA3 gene) [5,6] risk factors for AC. The Hosmer-Lemeshow test was used to verify goodness of fit of the model to the data. To determine whether a potential association with AH was independent of severity of the disease, we stratified patients into homogenous tertiles according to the model for end-stage liver disease (MELD) and used the Cochran-Mantel-Haenszel test. We explored the association of 25(OH)D with
346
mortality and complications of portal hypertension (ascites, encephalopathy, HRS, SBP, bleeding) in AC patients (n = 254) at one year. Follow-up time was defined as the period from the first day of medical therapy to the last followup visit. Patient survival, expressed in percentages, was estimated by the Kaplan-Meier method. The association between 25(OH)D levels and mortality/complications of portal hypertension was assessed by the log-rank test. With respect to the association with mortality, univariate, and multivariable (after adjustment for age, sex, BMI, seasonal timing of sample collection, and for the MELD) Cox proportional hazards models were performed. Data from in vitro and in vivo experiments are expressed as the median with the (minimum-maximum) values. All reported p-values are two-tailed. A p-value <0.05 was considered statistically significant. Statistical analyses were performed using SPSS version 19.0 software.
Results Association between 25(OH)D levels and clinical, biochemical, and histological parameters of ALD Healthy controls had a 25(OH)D median concentration of 19 ng/ ml, which was significantly higher than that of ALD patients (9.6 ng/ml) in univariate analysis (p = 4.75 1024, Table 1) and after adjustment for age, sex, BMI, and seasonal timing of sample collection (p = 1.45 108). After matching patients and healthy controls for age (±5 years) and sex (Supplementary Table 4) the difference in the 25(OH) median concentration remained highly significant in univariate (p = 3.46 1011) and multivariable analyses (p = 6.43 108). Clinical, biochemical and histological characteristics according to 25(OH)D levels (<10 ng/ml vs. P10 ng/ml) are summarized in Table 2. In addition, clinical, biological, and histological characteristics of cirrhotic and non-cirrhotic patients are reported in Supplementary Table 2. No association was observed between age, sex, BMI, and heavy alcohol consumption or 25(OH)D levels. Similarly, no significant association of 25(OH)D with ALT was seen, in contrast with AST (p = 1.00 103). Furthermore, patients with 25(OH)D below 10 ng/ml had a higher HVPG (p = 5.80 106). In addition, low 25(OH)D levels were strongly associated with severity of liver disease assessed by MELD (p = 2.50 104) and Child-Pugh scores (p = 8.50 107). Finally, a low vitamin D concentration was significantly associated with more severe steatosis (p = 1.00 103), fibrosis (p = 0.016), and AH (p = 2.14 104). After stratification of patients into tertiles according to the MELD score (<9; 9–16.2; >16.2), the association with AH remained statistically significant (p = 0.035). Conversely, we reported no association between vitamin D status and the Maddrey discriminant function (mDF) score in AH patients (p = 0.093), nor with the occurrence of HCC (p = 1.000). Remarkably, seasonal timing of sample collection for 25(OH)D measurement during summer and autumn was linked to higher levels compared to winter and spring months (p = 0.033). Association between 25(OH)D levels and cirrhosis and its related complications Low 25(OH)D concentrations were significantly associated with cirrhosis (odds ratio [OR] = 2.13, 95% confidence interval [CI] = 1.18–3.84, p = 0.013, Table 3) independently of other classical risk factors (age, sex, BMI, and heavy alcohol consumption) and seasonal timing of sample collection. In addition, we observed a significant link between cirrhosis and obesity (OR = 1.88, 95% CI = 1.04–3.40, p = 0.036, Table 3) and the
Journal of Hepatology 2013 vol. 59 j 344–350
JOURNAL OF HEPATOLOGY Table 2. Association of serum 25(OH)D levels with clinical, biological, and histological variables in ALD patients.
<10 ng/ml (n = 171) 38.0 53.4 ± 9.0 56.3 39.2 33.0 (25-55) 60 (43-107) 143.0 (59.3-323) 3.1 (1.2-9.3) 1.46 (1.12-1.78) 16 (12-21) 16.6/28.6/54.8 14.4 (8.8-21.0) 36.0 (23.8-59.0) 78.0 29.4 20 (9-50) 1.2/2.9/2.3/9.4/84.2 2.3
Female sex, % Age, yr BMI >25 kg/m2, % Blood sampling during summer/autumn, % ALT, IU/L AST, IU/L GGT, IU/L Bilirubin, mg/dl INR HVPG, mmHg Child-Pugh score A/B/C, % MELD score mDF Alcohol >80 g/day, % AH, % Steatosis, % Fibrosis stage 0/1/2/3/4, %* HCC, %
25(OH)D levels ≥10 ng/ml (n = 153) 32.0 55.4 ± 9.6 61.6 51.0 35 (24-49) 48 (34-78) 146.5 (72.0-297.0) 1.3 (0.7-2.7) 1.27 (1.08-1.54) 12 (6-17) 36.4/41.8/21.8 10.0 (6.4-15.4) 27.0 (7.8-45.8) 73.5 10.3 10 (5-35) 2/5.6/12.5/8.3/71.6 2.6
p value
0.260 0.057 0.339 0.033 0.874 1.00 x 10-3 0.816 1.42 x 10-7 3.20 x 10-4 5.80 x 10-6 8.50 x 10-7 2.50 x 10-4 0.071 0.356 7.01 x 10-5 1.00 x 10-3 0.016 1.000
25(OH)D, 25 hydroxyvitamin D; BMI, body mass index; ALT, alanine aminotransferase; AST, aspartate aminotransferase; GGT, gamma-glutamyl transferase; HCC, hepatocellular carcinoma; HVPG, hepatic venous pressure gradient; MELD, model for end-stage liver disease; AH, alcoholic hepatitis; mDF, Maddrey discriminant function test. Variables with normal distribution are presented as mean ± standard deviation. Skewed variables are presented as median (interquartile range). ⁄ Available in 97% of cases, including all non-cirrhotic patients.
Table 3. Univariate and multivariable analysis of risk of cirrhosis.
Variables Age Sex (female) BMI >25 kg/m2 Alcohol >80 g/day PNPLA3 rs738409 25(OH)D <10 ng/ml*
OR 1.02 1.14 2.15 1.35 2.24 1.73
Univariate 95% CI 0.99-1.05 0.65-2.00 1.45-3.57 0.74-2.46 1.31-3.83 1.01-2.95
p value 0.097 0.645 6.00 x 10-3 0.329 3.00 x 10-3 0.044
OR 1.03 1.39 1.88 1.71 2.42 2.13
Multivariable* 95% CI 0.99-1.06 0.75-2.56 1.04-3.40 0.89-3.29 1.36-4.30 1.18-3.84
p value 0.070 0.270 0.036 0.109 3.00 x 10-3 0.013
OR, odds ratio; CI, confidence interval; 25(OH)D, 25 hydroxyvitamin D; BMI, body mass index. ⁄ Adjusted for seasonal timing of sample collection.
rs738409 variant in the PNPLA3 gene (OR = 2.42, 95% CI = 1.36– 4.30, p = 3.00 103, Table 3). Allele frequencies for the PNPLA3 SNP were in Hardy-Weinberg equilibrium, indicating that genotype and allele frequencies were in equilibrium and suggesting the absence of a problem in genotyping (data not shown) [22]. The Hosmer–Lemeshow test was not significant (p = 0.644), suggesting good calibration of the model. Furthermore, a 25(OH)D concentration below 10 ng/ml was significantly associated with complications of portal hypertension at one year after adjustment for seasonal timing of sample collection and disease severity (MELD score in tertiles), including the presence of ascites (p = 1.22 104), encephalopathy (p = 5.22 103) and HRS (p = 0.025) (Supplementary Table 3). However, although SBP and bleeding were more prevalent in the low vitamin D concentration group, this difference was not statistically significant (p = 0.451 and p = 0.449, Supplementary Table 3). The proportion of patients lost to follow-up was nearly 50% for most events studied, but was not statistically different
between patients with 25(OH)D <10 and P10 ng/ml (data not shown). Association between 25(OH)D levels and mortality in AC patients The median follow-up was 4.9 months. After 12 months, 12.6% died (91% of this mortality was liver-related) and 11.4% underwent liver transplantation (Supplementary Table 2). During the same period, the proportion of patients lost to follow-up was 52% at 12 months. Nevertheless, this proportion was not statistically different in patients with or without severe vitamin D deficiency (48.3% vs. 57.9%, p = 0.129). A 25(OH)D concentration below 10 ng/ml was significantly linked to a lower overall survival rate at one year (hazards ratio [HR] = 5.95 95% CI = 2.09–16.96, p = 8.56 104, Fig. 1 and Table 4). In multivariable analysis, only the MELD score and severe 25(OH)D deficiency remained associated with one-year mortality in AC patients (HR = 4.33, 95% CI = 1.47–12.78, p = 7.94 103, Table 4).
Journal of Hepatology 2013 vol. 59 j 344–350
347
Research Article **
8000
0.9 TNFα (pg/ml)
p = 8.56 x 10-4
0.8 0.7
25(OH)D <10 ng/ml
*
HS AC
6000 4000 §§
§
2000
0.6
Vit D ≥10 112 Vit D <10 142
2 74 78
4 64 68
6 8 10 Follow-up (months) 58 63
51 55
45 52
12
14
43 45
Fig. 1. Kaplan Meier curves illustrating the probability of survival of alcoholic cirrhotic (AC) patients according to 25-hydroxyvitamin D [25(OH)D] cut-offs (<10 vs. P10 ng/ml). A total of 254 AC patients were followed for one year. The table below indicates the number of patients remaining at risk of mortality at each time point.
Influence of 1,25(OH)2D on TNFa production by stimulated PBMCs In PBMCs isolated from AC patients (n = 8; 25(OH)D = 10 [7– 60.2] ng/ml) and from healthy subjects (HS; n = 8; 25(OH)D = 27 [22–34] ng/ml), ConA stimulation induced a significant increase in TNFa production (HS, TNFa without ConA: 191.5 [0–614.8] pg/ml vs. TNFa after ConA: 5183.4 [1735.2–9859.8] pg/ ml, p = 3.00 103; AC, TNFa without ConA: not detectable vs. TNFa after ConA: 7487.7 [2377.5–12,000] pg/ml, p = 3.31 104), while 1 h pretreatment with 107 M 1,25(OH)2D significantly decreased TNFa production both in HS and in AC patients (HS, TNFa after 1,25(OH)2D + ConA: 1556.0 [873.2–3781.1] pg/ml vs. TNFa after ConA, p = 0.012; AC TNFa after 1,25(OH)2D + ConA: 1693.7 [1106.1–4314.7] pg/ml vs. TNFa after ConA, p = 3.00 103; Fig. 2). Influence of 1,25(OH)2D supplementation in a murine experimental model of ALD Serum levels of ALT and the liver/mouse weight ratio were both increased in the EtOH group compared to the Ctrl group, as expected from the experimental model (ALT: EtOH 50 [10– 100] IU/L vs. Ctrl 10 [10–32] IU/L, p = 1.40 104; liver/mouse weight ratio: EtOH 7.0 [6.2–8.1] vs. Ctrl 5.6 [4.8–6.3], p = 9.00 106). However, vitamin D supplementation had no significant impact on these variables (ALT: EtOH + 1,25(OH)2D 56 [36–132] pg/ml; liver/mouse weight ratio: EtOH + 1,25(OH)2D 7.1 [6.2–10.8], Fig. 3A and B). TNFa expression was induced in livers of alcohol-fed mice compared to the Ctrl group (TNFa expression in EtOH mice: 2.4
n.d.
0 ConA (2.5 µg/m l) 1,25(OH)2D (10-7M)
-
n.d.
+
+ -
+ +
Fig. 2. Effect of 1,25(OH)2D on TNFa production by stimulated PBMCs. PBMCs, isolated from patients with alcoholic cirrhosis (AC; n = 8) and healthy subjects (HS; n = 8), were pretreated with 107 M 1,25-dihydroxyvitamin D [1,25(OH)2D] or with vehicle control before stimulation with 2.5 lg/ml ConA. TNFa production was assayed in culture supernatants by ELISA. Bars represent median with interquartile range (duplicate cultures for each point). No stimulation vs. ConA stimulation, ⁄p = 3.00 103 in HS; ⁄⁄p = 3.31 104 in AC; ConA vs. 1,25(OH)2D + ConA; §p = 0.012 in HS; §§p = 3.00 103 in AC; n.d., not detectable.
A
Control EtOH EtOH + 1,25(OH)2D 80 *
§
60 ALT (IU/L)
0
40 20
B
Control EtOH EtOH + 1,25(OH)2D 10
Liver/mouse weight ratio
Survival probability
10,000
25(OH)D ≥10 ng/ml
1.0
8
*
§
6 4 2
0 Fig. 3. Effect of 1,25(OH)2D oral supplementation in a murine experimental model of ALD. (A) Plasma alanine aminotransferase (ALT) and (B) liver/mouse weight ratio assessment in mice fed the Lieber-DeCarli diet for 6 weeks. Eighteen mice from the ethanol group were orally administered 50 ng of 1,25(OH)2D three times a week for 5 weeks (EtOH + 1,25(OH)2D), while pair-fed (Ctrl; n = 7) and ethanol-fed (EtOH; n = 22) groups received PBS as control vehicle. Values are expressed as median and interquartile range. ALT, ⁄p = 1.40 104 EtOH vs. Ctrl; § p = 1.20 105 EtOH + 1,25(OH)2D vs. Ctrl; liver/mouse weight ratio, ⁄ p = 9.00 106 EtOH vs. Ctrl; §p = 2.40 105 EtOH + 1,25(OH)2D vs. Ctrl.
[1.0–7.1] times the expression in Ctrl mice, p = 1.00 103), while oral administration of 1,25(OH)2D significantly decreased expression of TNFa in the livers of EtOH-fed mice by 35% (p = 0.04, Fig. 4).
Table 4. Univariate and multivariable analysis of risk of mortality at one year.
Variables Age Sex (female) BMI >25 kg/m2 MELD score 25(OH)D <10 ng/ml*
HR 1.00 1.31 1.46 1.09 5.95
Univariate 95% CI 0.97-1.04 0.65-2.66 0.69-3.08 1.06-1.12 2.09-16.96
p value 0.837 0.451 0.323 1.62 x 10-9 8.56 x 10-4
HR 1.02 1.31 1.07 1.08 4.33
Multivariable* 95% CI 0.98-1.07 0.63-2.73 0.48-2.42 1.04-1.12 1.47-12.78
HR, hazards ratio; CI, confidence interval; 25(OH)D, 25 hydroxyvitamin D; BMI, body mass index; MELD, model for end-stage liver disease. ⁄ Adjusted for seasonal timing of sample collection.
348
Journal of Hepatology 2013 vol. 59 j 344–350
p value 0.295 0.471 0.864 1.15 x 10-6 7.94 x 10-3
JOURNAL OF HEPATOLOGY Control EtOH EtOH + 1,25(OH)2D Hepatic TNFα mRNA (fold increase)
4 * §
3
**
2 1 0
Fig. 4. Effect of 1,25(OH)2D oral supplementation on hepatic expression of TNFa after 6 weeks of the Lieber-DeCarli diet. Gene transcripts were analyzed by RT-PCR and results were normalized to HPRT mRNA levels using the 2DDCT method. Data are expressed as fold increase in normalized TNFa expression by livers of the pair-fed group (Ctrl; n = 7). Bars represent median with interquartile range. ⁄p = 1.00 103 EtOH vs. Ctrl; ⁄⁄p = 0.021 ethanol-fed mice orally supplemented with 1,25(OH)2D (EtOH + 1,25(OH)2D; n = 18) vs. Ctrl; §p = 0.04 ethanolfed mice (EtOH; n = 22) vs. EtOH + 1,25(OH)2D.
Discussion This study demonstrates for the first time that low serum levels of 25(OH)D are associated with increased liver damage (AST levels, steatosis, AH, and cirrhosis), and especially with more advanced liver disease (higher HVPG levels and MELD/Child-Pugh scores) in ALD patients. Furthermore, a low threshold (<10 ng/ml) was significantly linked to increased mortality. Low serum levels of vitamin D have long been observed in chronic liver diseases, especially at the cirrhosis stage [9]. They were also shown to be associated with liver damage, especially fibrosis, in NAFLD and CHC [11–13]. In line with these observations, we reported that low 25(OH)D concentrations were associated with cirrhosis and that the link remained significant even after adjustment for other classical risk factors (age, sex, alcohol consumption, and BMI), as was the PNPLA3 (rs738409 C>G) variant [4–6]. Moreover, like previous authors, we observed that the vitamin D concentration was closely related to liver disease severity as assessed by Child-Pugh [23], and MELD scores in AC patients. We also reported a novel link to AH, portal hypertension, and most of its related complications at one year. Interestingly, the association with AH was independent of severity of the disease. In line with the previous considerations, the link between higher mortality and severe vitamin D deficiency was therefore expected. Thus, a low 25(OH)D concentration remained associated with mortality independently of the MELD score and was the leading risk factor associated with short-term mortality in liver cirrhosis, suggesting that severe 25(OH)D deficiency is a marker of AC severity/prognosis. Generally speaking, the proportion of patients lost to follow-up was substantial (nearly 50% at one year for most events studied). This limitation prevented an accurate analysis over a longer period of time. Nonetheless, the percentage of lost to follow-up patients was not statistically different between patients with severe 25(OH)D deficiency and those without. Overall, although these associations are interesting, further studies using a prospective design are required to confirm these findings. In addition, severe vitamin D deficiency was not associated with HCC; however, this analysis again requires prospective cohorts of AC followed for a much longer time to reach a firm conclusion.
Sources of vitamin D include increased production through sunlight and, to a much lower extent, diet and dietary supplements. Nutritional status was not assessed in this study, which may be considered another limitation. Nevertheless, adequate assessment of malnutrition is particularly challenging in ALD [24]. Moreover, the effect of dietary intake on 25(OH)D status is relatively limited, since few foods contain sizable amounts of vitamin D. Conversely, the vast majority of 25(OH)D is derived from the skin through sunlight exposure [9]. Thus, based on observations in NAFLD and CHC, we also report that seasonal timing of the sampling had a significant influence on vitamin D status in ALD patients [10,12]. Overall, these association data using a cross-sectional design do not permit establishing a causal relationship between vitamin D severe deficiency and ALD severity/mortality. However, growing evidence suggests that 25(OH)D might play a biological role in various liver diseases [10,11,13]. The increased levels of TNFa found in ALD patients are associated with development of complications and with the severity and mortality of cirrhosis [7]. Indeed, a leading theory in ALD pathogenesis suggests the deleterious consequences of a pro-inflammatory cascade, favored by ethanol and followed by a dysregulated immune response to stimulation [25]. Experimental studies have shown that 1,25(OH)2D can suppress the release of TNFa both in vivo [26,27] and in vitro [28]. Interestingly, our in vitro experiments confirmed the inhibitory effect of 1,25(OH)2D on TNFa production and suggest that this inhibitory mechanism is not altered by hepatic disease. Furthermore, in our in vivo experiments, 1,25(OH)2D oral supplementation reduced hepatic expression of TNFa induced by an ethanol diet. Although correction of vitamin D status did not modify transaminase serum levels, these results, in line with in vitro human data on PBMCs, suggest that it may downplay the deleterious inflammatory cascade that takes place in the liver following excessive alcohol consumption, by reactivating the functional inhibitory pathway. This hypothesis is further supported by a recently published study demonstrating that vitamin D deficiency exacerbates the development of NAFLD in rats through increased expression of inflammatory genes in the liver [29]. In conclusion, this study reports that 25(OH)D levels are significantly associated with ALD severity, portal hypertension, its related complications and mortality at one year. In addition, our data suggest that 1,25(OH)2D could improve the deleterious pro-inflammatory cytokine profile of AC patients. Although our results must be replicated in prospective cohorts to address the true impact of 25(OH)D severe deficiency on the ALD phenotype, vitamin D may well represent both a biomarker and a therapeutic target worth exploring in ALD. Financial support This work was supported by the Fonds Erasme for Medical Research and the FNRS (Belgian National Fund of Scientific Research).
Conflict of interest The authors who have taken part in this study declared that they do not have anything to disclose regarding funding or conflict of interest with respect to this manuscript.
Journal of Hepatology 2013 vol. 59 j 344–350
349
Research Article Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jhep.2013. 03.024.
[14]
[15] [16]
References [1] Rehm J, Mathers C, Popova S, Thavorncharoensap M, Teerawattananon Y, Patra J. Global burden of disease and injury and economic cost attributable to alcohol use and alcohol-use disorders. Lancet 2009;373:2223–2233. [2] European, Association for the Study of the Liver. EASL clinical practical guidelines: management of alcoholic liver disease. J Hepatol 2012;57:399–420. [3] Rehm J, Taylor B, Mohapatra S, Irving H, Baliunas D, Patra J, et al. Alcohol as a risk factor for liver cirrhosis: a systematic review and meta-analysis. Drug Alcohol Rev 2010;29:437–445. [4] Naveau S, Giraud V, Borotto E, Aubert A, Capron F, Chaput JC. Excess weight risk factor for alcoholic liver disease. Hepatology 1997;25:108–111. [5] Stickel F, Buch S, Lau K, Zu Schwabedissen HM, Berg T, Ridinger M, et al. Genetic variation in the PNPLA3 gene is associated with alcoholic liver injury in caucasians. Hepatology 2011;53:86–95. [6] Trépo E, Gustot T, Degre D, Lemmers A, Verset L, Demetter P, et al. Common polymorphism in the PNPLA3/adiponutrin gene confers higher risk of cirrhosis and liver damage in alcoholic liver disease. J Hepatol 2011;55:906–912. [7] Gao B, Bataller R. Alcoholic liver disease: pathogenesis and new therapeutic targets. Gastroenterology 2011;141:1572–1585. [8] Holick MF. Vitamin D deficiency. N Engl J Med 2007;357:266–281. [9] Kitson MT, Roberts SK. D-livering the message: the importance of vitamin D status in chronic liver disease. J Hepatol 2012;57:897–909. [10] Targher G, Bertolini L, Scala L, Cigolini M, Zenari L, Falezza G, et al. Associations between serum 25-hydroxyvitamin D3 concentrations and liver histology in patients with non-alcoholic fatty liver disease. Nutr Metab Cardiovasc Dis 2007;17:517–524. [11] Petta S, Camma C, Scazzone C, Tripodo C, Di Marco V, Bono A, et al. Low vitamin D serum level is related to severe fibrosis and low responsiveness to interferon-based therapy in genotype 1 chronic hepatitis C. Hepatology 2010;51:1158–1167. [12] Bitetto D, Fattovich G, Fabris C, Ceriani E, Falleti E, Fornasiere E, et al. Complementary role of vitamin D deficiency and the interleukin-28B rs12979860 C/T polymorphism in predicting antiviral response in chronic hepatitis C. Hepatology 2011;53:1118–1126. [13] Lange CM, Bojunga J, Ramos-Lopez E, von Wagner M, Hassler A, Vermehren J, et al. Vitamin D deficiency and a CYP27B1-1260 promoter polymorphism are
350
[17]
[18]
[19]
[20]
[21]
[22] [23]
[24] [25] [26]
[27] [28]
[29]
associated with chronic hepatitis C and poor response to interferon-alfa based therapy. J Hepatol 2011;54:887–893. Faria SC, Ganesan K, Mwangi I, Shiehmorteza M, Viamonte B, Mazhar S, et al. MR imaging of liver fibrosis: current state of the art. Radiographics 2009;29:1615–1635. Lucey MR, Mathurin P, Morgan TR. Alcoholic hepatitis. N Engl J Med 2009;360:2758–2769. European, Association for the Study of the Liver, European, Organisation for Research and Treatment of Cancer. EASL-EORTC clinical practice guidelines: management of hepatocellular carcinoma. J Hepatol 2012;56: 908–943. European, Association for the Study of the Liver. EASL clinical practice guidelines on the management of ascites, spontaneous bacterial peritonitis, and hepatorenal syndrome in cirrhosis. J Hepatol 2010;53: 397–417. Brunt EM, Janney CG, Di Bisceglie AM, Neuschwander-Tetri BA, Bacon BR. Nonalcoholic steatohepatitis: a proposal for grading and staging the histological lesions. Am J Gastroenterol 1999;94:2467–2474. Bosch J, Abraldes JG, Berzigotti A, Garcia-Pagan JC. The clinical use of HVPG measurements in chronic liver disease. Nat Rev Gastroenterol Hepatol 2009;6:573–582. McMullen MR, Pritchard MT, Wang Q, Millward CA, Croniger CM, Nagy LE. Early growth response-1 transcription factor is essential for ethanol-induced fatty liver injury in mice. Gastroenterology 2005;128:2066–2076. Bischoff-Ferrari HA, Giovannucci E, Willett WC, Dietrich T, Dawson-Hughes B. Estimation of optimal serum concentrations of 25-hydroxyvitamin D for multiple health outcomes. Am J Clin Nutr 2006;84:18–28. Balding DJ. A tutorial on statistical methods for population association studies. Nat Rev Genet 2006;7:781–791. Malham M, Jorgensen SP, Ott P, Agnholt J, Vilstrup H, Borre M, et al. Vitamin D deficiency in cirrhosis relates to liver dysfunction rather than aetiology. World J Gastroenterol 2011;17:922–925. McClain CJ, Barve SS, Barve A, Marsano L. Alcoholic liver disease and malnutrition. Alcohol Clin Exp Res 2011;35:815–820. Tsukamoto H, Lu SC. Current concepts in the pathogenesis of alcoholic liver injury. FASEB J 2001;15:1335–1349. Zhu Y, Mahon BD, Froicu M, Cantorna MT. Calcium and 1 alpha,25dihydroxyvitamin D3 target the TNF-alpha pathway to suppress experimental inflammatory bowel disease. Eur J Immunol 2005;35:217–224. Lemire JM. Immunomodulatory role of 1,25-dihydroxyvitamin D3. J Cell Biochem 1992;49:26–31. Muller K, Haahr PM, Diamant M, Rieneck K, Kharazmi A, Bendtzen K. 1,25Dihydroxyvitamin D3 inhibits cytokine production by human blood monocytes at the post-transcriptional level. Cytokine 1992;4:506–512. Roth CL, Elfers CT, Figlewicz DP, Melhorn SJ, Morton GJ, Hoofnagle A, et al. Vitamin D deficiency in obese rats exacerbates nonalcoholic fatty liver disease and increases hepatic resistin and Toll-like receptor activation. Hepatology 2012;55:1103–1111.
Journal of Hepatology 2013 vol. 59 j 344–350